OVM-OSN: an optimal validation model applied to detection of fake accounts on online social networks

Abstract

Social network sites have millions of users, all sharing personal information in an unwilling manner with friends and contacts. The recent reports point out these networks are spread-through with millions of fake accounts, which affects the users' security and privacy. To overcome such issues, online social networks (OSNs) utilise the fake detection methods to preserve the user privacy and system reliability. Since, the fake account detection is very predominant and a crucial process in OSNs, in this paper, we propose an optimal validation model (OVM), which detects the fake accounts on OSNs, named as OVM-OSN. It is a simple yet efficient computational process for community detection. In OVM-OSN, we employ a novel community detection method utilising the multi-swarm fruit fly optimisation. Then, we use fuzzy-based decision model to differentiate the fake from normal accounts, which maximise the trustworthiness of online identities. Hence, this proposed OVM-OSN method is reliable even under link, node failure strategies and it is tested with Facebook and Google+ networks. Simulation results show the effectiveness of OVM-OSN method in terms of detection rate compared to the cooperative and adaptive decentralised identity validation model.